Pre-training strategies and datasets for facial representation learning

TL;DR

This paper evaluates pre-training strategies and datasets for facial representation learning, introducing a comprehensive benchmark and demonstrating that unsupervised pre-training on uncurated data improves performance across multiple facial analysis tasks.

Contribution

It introduces a new benchmark for facial representation learning, systematically compares supervised and unsupervised pre-training, and analyzes dataset properties affecting facial task performance.

Findings

Unsupervised pre-training on uncurated data improves facial task accuracy.

Many facial video datasets contain significant redundancy.

The study provides extensive experimental evidence supporting unsupervised learning benefits.

Abstract

What is the best way to learn a universal face representation? Recent work on Deep Learning in the area of face analysis has focused on supervised learning for specific tasks of interest (e.g. face recognition, facial landmark localization etc.) but has overlooked the overarching question of how to find a facial representation that can be readily adapted to several facial analysis tasks and datasets. To this end, we make the following 4 contributions: (a) we introduce, for the first time, a comprehensive evaluation benchmark for facial representation learning consisting of 5 important face analysis tasks. (b) We systematically investigate two ways of large-scale representation learning applied to faces: supervised and unsupervised pre-training. Importantly, we focus our evaluations on the case of few-shot facial learning. (c) We investigate important properties of the training datasets including their size and quality (labelled, unlabelled or even uncurated). (d) To draw our conclusions, we conducted a very large number of experiments. Our main two findings are: (1) Unsupervised pre-training on completely in-the-wild, uncurated data provides consistent and, in some cases, significant accuracy improvements for all facial tasks considered. (2) Many existing facial video datasets seem to have a large amount of redundancy. We will release code, and pre-trained models to facilitate future research.